Prognostic Implications of the Detection of Measurable Residual Disease and Mutations Based on Next‐generation Sequencing in Acute Myeloid Leukaemia with Biallelic Mutations of CEBPA
Jing Wang,Hao Jiang,Jie Yu,Long Su,Daihong Liu,Daobin Zhou,Jingbo Wang,Hui Liu,Guo‐Rui Ruan,Sujun Gao,Xiao Jun Huang
DOI: https://doi.org/10.1111/bjh.18143
2022-01-01
British Journal of Haematology
Abstract:Mutations in the CCAAT/enhancer binding protein α (CEBPA) gene occur in approximately 10% of all acute myeloid leukaemia (AML) cases, where half of the mutations are biallelic CEBPA (bi CEBPA), which has now been acknowledged to be associated with a favourable prognosis.1-5 However, long-term follow-up of cytogenetically normal CEBPA-mutated AML showed that the 5-year and 10-year cumulative incidence of relapse (CIR) rates were 37%–40% and 45% respectively, in bi CEBPA AML patients.4, 6 Therefore, the rapid identification of high-risk relapse patients and pre-emptively treating patients with more aggressive therapies may improve survival. Between January 2012 and April 2020, 262 newly diagnosed bi CEBPA de novo AML patients were enrolled in six Chinese centres. A total of 149 patients (56.9%) were men. The median age of the patients at diagnosis was 40 years (range 16–79 years). Among the 260 patients who achieved complete remission (CR), 224 received intermediate-dose cytarabine-based chemotherapy and 36 received chemotherapy followed by allogeneic- haematopoietic stem cell transplantation (allo-HSCT) in first complete remission (CR1; human leukocyte antigen-identical sibling donor, n = 9; matched unrelated donor, n = 3; haploidentical related donor, n = 24) as post-remission therapy. Twenty-nine patients who relapsed after chemotherapy received allo-HSCT as salvage therapy. One hundred and twenty-four of the patients at Peking University have been reported previously.7 All patients provided informed consent, according to the procedures of the institutional review board of each centre and in accordance with the Declaration of Helsinki. The cut-off date for follow-up was 28 February 2021. Genomic DNA extraction, targeted gene list, quality control data, access number gene panel, targeted sequencing panel data, and sequencing depth definition have been described in detail.7 Survival functions were estimated using the Kaplan–Meier method and compared using the log-rank test. Competing risk analyses were used to calculate the CIR and non-relapse mortality using Grey's test to examine differences between the groups. Parameters with a p value of < 0.2 by univariate analysis, were entered into a multivariate analysis using a Cox proportional hazards model, to identify the statistically significant parameters associated with survival, and the Fine-Grey model for CIR. Data analyses were performed using SPSS, version 24.0 software (IBM Corporation) and GraphPad Prism 6 (GraphPad Software Inc.). R statistical software version 3.6.1 (http://cran.R-project.org) was used for the competing-risks analysis. With a median follow-up of 34 months (range 0.5–109) in all patients and 43.5 months (range 10–109 ) in the 209 survivors, 88 patients relapsed at the median follow-up of 10.5 months (range 3–38). At the 5-year follow-up, the probabilities of CIR, relapse-free survival (RFS), and overall survival (OS) were 38.0%, 58.8%, and 75.7% respectively [with a 95% confidence interval (CI) of 34.7–41.3, 55.5–62.1, and 72.6–78.8 respectively]. Of these, 257 patients were analysed further by a sensitive next-generation sequencing (NGS) assay for mutations in 87 candidate genes and target regions. The cohort from the earlier study is seen as the training cohort (n = 124),7 the extra patients included in this study serve as the confirmation cohort (n = 138), the patient characteristics are comparable and shown in Table 1. Forty-nine other molecular mutations were detected in the patients with bi CEBPA. Sixty-two patients (24.1%) had no additional molecular mutation, while 100 (38.9%), 50 (19.5%), 26 (10.1%), 10 (3.9%), two (0.8%), and seven (2.7%) had one, two, three, four, five, and six or more additional mutations, respectively. The most common co-occurring mutations were found in the GATA2 (n = 64/257, 24.9%), WT1 (n = 57/257, 22.3%), CSF3R (n = 40/257, 15.6%), NRAS (n = 37/257, 14.4%), and TET2 (n = 28/257, 10.9%) genes. Without additional molecular mutations, the GATA2 mutation had good trends for lower CIR and longer RFS, the WT1 mutation had a poor factor for 5-year CIR (49.6% vs. 33.3%, p = 0.011), and 5-year RFS (48.1% vs. 63.5%; p = 0.017), and the CSF3R mutation had an inferior trend for higher CIR and shorter RFS. Patients without additional molecular mutations and with a GATA2 mutation were defined as “NGS low risk” (n = 92, 40.4%), while those with other co-occurring mutations were defined as “NGS high risk” (n = 136, 59.6%). Low-risk NGS had a favourable factor for 5-year CIR (28.8% vs. 43.0.8%, p = 0.042), 5-year RFS (69.4% vs. 53.4%; p = 0.021), and 5-year OS (82.2% vs. 68.9%, p = 0.026) (Figure 1). Minimal residual disease (MRD) was negative at the median follow-up of 2 months (range 1–7) Patients with positive MRD status after two consolidation cycles, and MRD-negative status lost at any time, had a higher CIR (5-year CIR, 62.3% vs. 31.0%; p < 0.001; 79.3% vs. 13.7%; p < 0.001 respectively) and worse RFS and OS (5-year RFS, 33.9% vs. 66.0%; p <0.001; 19.5% vs. 82.0%; p < 0.001 respectively; 5-year OS, 61.0% vs. 80.5%; p = 0.002; 61.1% vs. 88.5%; p < 0.001 respectively) than those with persistent negative MRD status. Patients with positive MRD status after two consolidation cycles and/or MRD-negative status lost at any time, were defined as “MRD high risk”, while those with persistent negative MRD status were defined as “MRD low risk.” Low-risk MRD was a favourable factor for 5-year CIR (13.4% vs. 83.9%, p < 0.001), 5-year RFS, (82.1% vs. 14.9%, p < 0.001), and 5-year OS (88.6% vs. 52.8%, p < 0.001) (Figure 1). Allo-HSCT after CR1 was a favourable factor for CIR (5-year CIR 6.8% vs. 42.4%; p < 0.001) and RFS (5-year RFS 81.6% vs. 55.6%, p = 0.010). For the relapse patients who obtained molecular remission during therapy, the median time was 3.5 months (range 0–22) from molecular relapse to clinical relapse. Patients with persistent positive MRD status after two consolidation cycles, and/or MRD-negative status lost at any time (molecular relapse), were redefined as high-risk patients and recommended for allo-HSCT as pre-emptive treatment. Thirteen molecular relapse patients who underwent allo-HSCT in CR1 as pre-emptive treatment were excluded in multivariate analysis in case of bias. Multivariate analysis of pre-treatment parameters showed that a low risk of NGS with a white blood cell count (WBC) of <20 × 109/L at diagnosis were favourable factors for CIR [hazard ratio (HR) = 0.67, 95% CI: 0.41–0.92, p = 0.048; HR = 0.54, 95% CI: 0.34–0.84, p = 0.002 respectively] and RFS (HR = 1.52, 95% CI: 1.01–2.47, p = 0.045; HR = 1.87, 95% CI: 1.19–2.91, p = 0.006 respectively), while a low risk of NGS was a favourable factor for OS (HR = 1.87, 95% CI: 1.11–3.67, p = 0.031). Multivariate analysis including pre-treatment and post-treatment parameters showed that a low risk of MRD was the only favourable factor for CIR (HR = 0.10, 95% CI: 0.06–0.18, p < 0.001), RFS (HR = 7.28, 95% CI: 4.48–11.78, p < 0.001), and OS (HR = 5.58, 95% CI: 2.84–10.73, p < 0.001) (Table 2). In 88 relapsed patients, 41 out of 76 patients (53.9%) achieved CR2 after induction. Nineteen CR2 and 10 non-remission patients underwent allo-HSCT and achieved a median survival time of 55 months (range 12–109) after relapse. Meanwhile, another 47 patients underwent a non-transplant treatment strategy as soon as relapse achieved a median survival time of 19 months (range 9–76) after relapse. The CR2 (5-year OS 89.2% vs. 26.4%; p = 0.001) and not reached (NR) (3-year OS, 67.5% vs. 19.4%; p = 0.009) patients after relapse who underwent allo-HSCT all achieved superior OS compared to non-transplant patients (Figure 2). The patients who underwent allo-HSCT, or not, after relapse were comparable with abnormal karyotype, FLT3-ITD mutation, or not, and age. In accordance with our previous study,7 we confirmed that combining pre- and post-treatment parameters is a rational way to improve outcome prediction in bi CEBPA AML. The results indicate that co-occurring mutations with a WBC count of >20 × 109/L at diagnosis in bi CEBPA have prognostic significance in the analysis of pre-treatment parameters. However, they were not the strongest factor predictors, especially compared with MRD risk stratification, when analysis of all factors included pre- and post-treatment. As with previous studies,1, 7 our results show that no OS benefits from allo-HSCT in CR1. However, 19 CR2 and 10 NR patients who underwent allo-HSCT achieved superior OS, calculated after relapse compared to non-transplant patients. The results are consistent with those of a European multicentre report8 and our previous single-centre report.7 This suggests that bi CEBPA mutations are predictive of a favourable outcome, and relapsed patients still have a favourable outcome followed by allo-HSCT. A limitation of our current study is that the patients were not randomised to receive chemotherapy or allo-HSCT. The patients who underwent allo-HSCT in CR1 were carrying a more abnormal karyotype, were more FLT3-ITD positive, and were younger than the non-transplant cohort. In summary, this multicentre study included a larger sample size and a longer follow-up period than our previous single-centre report.7 Patients without additional molecular mutations and with a GATA2 mutation were defined as “NGS low risk”, which was a favourable factor of CIR, RFS, and OS. Meanwhile, a WBC count of <20 × 109/L at diagnosis was a favourable factor of CIR and RFS in the analysis of pre-treatment parameters. A high risk of MRD was the only adverse factor of relapse and survival when all factors included pre-treatment and post-treatment. Relapsed patients still have a favourable outcome after reinduction, followed by allo-HSCT. These conclusions need further verification through prospective studies. The work was supported by "the National Key Research and Development Program of China" (2017YFA0104500), "the National Natural Science Foundation of China" (grant no. 81930004), "Innovative Research Groups of the National Natural Science Foundation of China" (grant no. 81621001), and "Capital Characteristic Clinic Project Foundation" (grant no. Z181100001718126).